Speech signal blind separating method based on variable step size natural gradient algorithm

A natural gradient algorithm and voice signal technology, applied in voice analysis, instruments, etc., can solve the problems of unrecognizable noise, uncompleted voice signal, distortion, etc., and achieve fast separation speed, accurate and stable separation effect

CN103903631AActive Publication Date: 2014-07-02HARBIN ENG UNIV
7 Cites 4 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Publication Date
2014-07-02

Smart Images

  • Figure 1
    Figure 1
  • Figure 2
    Figure 2
  • Figure 3
    Figure 3
Patent Text Reader

Abstract

The invention provides a speech signal blind separating method based on a variable step size natural gradient algorithm. (1) A speech signal collecting device is used for collecting mixed speech singles of speeches of multiple speakers, and the number of microphones is larger than or equal to that of the speakers; (2) the collected mixed speech signals are preprocessed, and then mixed speech signals which have zero mean, irrelevant signal sources and the high signal to noise ratio are obtained, wherein the preprocessing comprises the steps of adopting an LMS digital filter, removing the mean and removing related whitening processing; (3) the estimation of the speech of each speech source is restored and obtained from the mixed human speeches by using the variable step size natural gradient algorithm for regulating step size based on gradient factors. The speech signal blind separating method based on the variable step size natural gradient algorithm is capable of separating real mixed speech signals, high in separating speed and accurate and stable in separating effect.
Need to check novelty before this filing date? Find Prior Art

Description

technical field

[0001] The present invention relates to a speech signal processing method, in particular to a blind separation algorithm of multi-sound source mixed signal variable step size natural gradient, and a separation system of mixed speech signals obtained thereby. Background technique

[0002] Blind source separation is an emerging research field that developed rapidly at the end of the 20th century. As a new data processing method, it is the product of the combination of artificial neural network, statistical signal processing, information theory, and computer, and has become an important part of some of the above fields. It has played an important role in the important topics of development and development, especially in the applications of biomedicine, speech signal processing, image processing, remote sensing, radar and communication systems.

[0003] In the field of speech signal processing, the current speech recognition and noise reduction enhancement algori...

Examples

Embodiment Construction

[0018] The following examples describe the present invention in more detail.

[0019] 1. Acquisition of voice mixed signal

[0020] According to the sampling theorem: the sampling frequency should be greater than or equal to twice the maximum frequency of the original signal. The frequency range of voice is 0~4kHz, so the minimum sampling frequency for voice signal is 8kHz, so the distance between any two microphones should satisfy where c is the speed of sound in air, f max =4kHz is the maximum frequency of the voice signal. In the process of collecting voice signals, the spatial position of the microphones is placed arbitrarily, but the distance between any two microphones is greater than 4.25cm. The collected analog voice signals are converted into digital voice signals through 8kHz sampling frequency. The digital signal of the i-th microphone is m i =[m i (1),...,m i (N)], N is the number of sampling points of the signal, and the signals collected by all microphones...